Download What Is a Disaster?

CHAPTER 1
What Is a Disaster?
Ronald W. Perry
Fieldwork is stimulating, challenging, and provides immediate rewards for the researcher.
Although contemplating theoretical and paradigmatic issues in one’s office may be less exciting
by comparison, it is important to deal with such tasks. Devising a definition of disasters or
assessing consensus on a definition is not only a part of sound theory and methodology (Bunge,
1998) but also contributes to a clearer vision of the field of study, and on a very practical level,
helps to sort out apparent anomalies in research findings and sets the stage for a progression
from simple description toward the social scientific tasks of explanation, prediction, and control
(Homans, 1967).
This chapter does not propose a new or unique definition of disasters, but rather recounts
efforts to define disasters by social scientists, particularly sociologists. This is accomplished
in several phases. First, attention is given to issues associated with definitions, including
clarifying the goal of defining disasters and the type of definition of interest. The task of
presenting definitions from the literature is tackled next. Finally, the definitions are reviewed
to assess levels of consensus and the presence of common themes.
WHAT KIND OF DEFINITION?
Seeking or proposing definitions of disaster can be a complex task that brings out the pedantic
in scholars and may create considerable frustration (Cutter, 2005a). Some of the complexity
and frustration can be addressed by specifying the purpose and audience for definitions of
disasters. Such definitions must be placed into a meaningful context that clarifies the essential
goal of the definition and the uses to which the definition is to be put. At the outset, it must be
acknowledged that the goals in creating definitions vary and that there is no single legitimate
purpose or content for definitions. Further, one must clarify whether disaster is being defined
as a concept or as an area of study, although there is an inevitable overlap between the two
approaches.
To attack the latter issue first, concern in this chapter is with the definition of disaster less
as a concept than as an area of study. Of course, the two ideas are not completely separable
and they clearly overlap. Certainly for methodologists and philosophers of science, the term
concept has a very specific meaning in the theoretical lexicon. However, while defining an
area of study has implications for theory and theory construction, the direct aim is more
1
2
Ronald W. Perry
meta-theoretical in that one seeks to introduce parameters on what is to be studied. At this
stage, one can avoid becoming immersed in the challenge of creating nominal and operational
definitions that pertain largely to concepts and the conceptualization process.
Hempel (1952) makes a useful distinction between real and nominal definitions. A real
definition, also called a connotative definition (Cohen, 1980, p. 143), is a statement that specifies or identifies the critical properties or features of the concept that is being defined. For
Hempel, this type of definition is in effect a class term intended to capture—with a degree of
openness or ambiguity—phenomena within an umbrella of meaning. The example he uses is a
chair defined as “. . . a separate movable seat for one person” (Hempel, 1952, p. 2). On the other
hand, a nominal definition may be seen as an expression of detailed characteristics that are tied
to a given term, which usually represents a given concept. Zetterberg (1965) emphasizes the
inductive nature of nominally defining a concept in his example that low levels of opportunity,
substandard housing, and deficient medical care are observables that reflect the term poor,
and are captured in the concept of poverty. The nominal definition forms the “meaning framework” for a concept that is scrutinized when developing operational definitions to initiate an
inductive or deductive research process, and research is ultimately aimed at or used for theory
construction.
The notion of real or connotative definition leads down a different path, one more dependent on the philosophy of science—whether one emphasizes a positivist, modified positivist,
or non-positivist approach (Martindale, 1979, p. 21). While there has been much criticism
(Masterman, 1970) and revision (Kuhn, 1970; Ritzer, 1979) of the sociological use of paradigm
over the years, it remains a useful—if still loosely used—idea. Thus, Ritzer (1979, p. 26) sees a
paradigm as the most fundamental picture of scientific subject matter, as the feature that defines
“what should be studied, what questions should be asked. . . ” From this perspective, defining
an area of study overlaps the problem of identifying a paradigm. Equally important, paradigms
and areas of study are consensus based, and providing definitions is not an empirical task but
an intellectual exercise resulting in an abstract construction (Kaplan, 1964). What is sought,
in the context of this chapter, are definitions of disaster that address concerns of paradigm and
do so by identifying critical features or characteristics of disasters.
WHO DOES THE DEFINING?
This discussion prefaces another distinction in defining disaster. Who has the “right” to propose
such definitions? In reality anyone has the right to propose a definition of disaster, and the
definition proposed depends on the purposes or interests of the definer. Kroll-Smith and Gunter
(1998) embrace what they call an interpretive voice and emphasize the notion that sociologists
should look for the definition of disaster among those who experience it (and are studied
by sociologists). Buckle (2005) notes that government develops “mandated” definitions of
disaster to determine the boundaries of emergency management and response; in the United
States, Presidential Disaster Declarations use these types of definitions. Britton (1986b) argues
that emergency managers have a specific perspective on what constitutes a disaster and are
often forced to simultaneously deal with definitions that differ between levels of government
and between specific policy audiences. Shaluf, Ahmadun, and Mustapha (2003) describe the
role of regulatory agencies in defining technological disasters. Others who propose and use
definitions of disaster include journalists, historians, and social scientists.
Quarantelli (1987b) has argued that there is no basis in logic and little hope in practice
that a single definition can be devised that meets and is universally accepted and useful.
What Is a Disaster?
3
Indeed, “heart attack” may convey to a victim all that he or she needs to know and at the
same time be only a vague description of an ailment to a cardiologist. It is necessary to
recognize that disaster will always mean many things to many people, and the description will
serve many different purposes—thus there will be many definitions. What becomes important
is the specification of the audience for the definition, bearing in mind the use to which that
audience will put the definition. Quarantelli (2005a) emphasizes that as social scientists—
sociologists in particular—defining disasters, we need to devote attention to the sociological
context and tradition, attending in particular to delimiting the phenomenon to become a focus
for the processes of social science. This chapter follows Quarantelli’s admonition.
The definition of disaster of interest here is one to be used by social scientists to delineate an
area of study and in so doing set the stage for knowledge accumulation and theory construction.
This is not to say that citizen perceptions of disasters—or the definition of disasters generated
by any other collectivity for that matter—are less important. All are entitled to their definition
and each is legitimate and most likely serves an intended purpose. For sociologists, the content
and patterns of such definitions are even a reasonable focus of research. However, the goal
here is to deal with definitions of disaster proposed by social scientists for social scientific
purposes.
THE CONTEXT OF DEFINITION ISSUES
Even when the type of definition, its purpose, and audience are specified, a challenge remains
in devising—or recounting—definitions of disaster. One issue is that, even if we limit scrutiny
to social science, several definitions are available at any one time, not to mention a large
number of empirical studies—some executed with an explicit definition in mind, most not—
with which to contend. Thus, when one proposes a definition of disaster, it may be an abstract
and nonempirical exercise, but there is certainly reason to reflect on previous definitions and
research. Social science cannot be conducted in an intellectual and empirical vacuum. If one
assumes that the definitions were proposed and the research carried out in good faith and with
professionalism, then each represents a legitimate attempt to either capture the meaning or
operate within the meaning of disaster. Consequently, prior definitions and studies at least
have the potential to inform current visions for definitions.
The challenge in using this information rests in the diversity of expression as well as within
the changing contexts in which disaster research has been undertaken. There is some consensus
that Samuel Prince’s (1920) dissertation on the Halifax explosion was the first systematic study
of disaster. A decade later, Carr (1932) addressed issues of substance, definition, and sequence
in disasters. In so doing, Carr was the first to describe disasters as inherently rooted in social
change. However, the real growth in disaster studies began in the early 1950s, accelerated
with the founding of the Disaster Research Center in 1963, and the field has virtually exploded
since the mid-1970s (Tierney, Lindell, & Perry, 2001). Indeed, in his seminal review of disaster
findings in 1986, Drabek found about one thousand empirical studies and the rate of research
has expanded since then. Interestingly, only a very small number of these researchers dealt
much with the definition of disasters. In fact, defining disasters became a widespread concern
only since the publication of Quarantelli’s (1987b) Presidential Address to the International
Research Committee on Disasters and much of that attention is testimony to Quarantelli’s
perseverance.
In the early decades of disaster research, definitions of the phenomenon were commonly
left implicit or partial, a state of affairs observed not just in disaster research or among
4
Ronald W. Perry
sociologists, or in the social sciences for that matter. For example, Carr identified a disaster as a product of its consequences, arguing that if the walls withstand the earthquake and
the dam retains the water, there is no disaster. Instead, he looks at disaster as the “collapse
of the cultural protections” (Carr, 1932, p. 211). The implicit definition that a disaster is any
event that generates significant negative consequences seems to have resulted in identification
of disasters with events in the natural environment (floods, earthquakes, severe storms, etc.),
technological incidents, and wartime incidents (Dombrowsky, 1981). This “disaster as negative, agent-caused event” approach can still be found in spite of early work distinguishing
disasters from other events (civil disturbances and wars, for example) associated with negative consequences (Barton, 1963; Quarantelli, 1966; Warheit, 1972). Quarantelli (1982b) was
among the first scholars to aggressively question this practice of defining disasters by surface
characteristics of the agent. The early 1960s saw a formally proposed social scientific definition
of disasters by Charles E. Fritz, first in a chapter on disasters in a social problems textbook
(1961a) and subsequently in a social science encyclopedia (1968). These definitional efforts
were followed closely by Barton’s seminal examinations of disasters and creation of a typology
in 1963.
The point here is that when one proposes a definition of disasters, one does not start from
scratch. As much as it might be appealing to focus solely on the intellectual abstract task, we
are influenced by, and need to acknowledge, our reading of the literature. After all, definitions
are largely the product of an inductive process. Often, this involves looking backwards and
making inferences to classify rather than eliminate research, while at same time exercising
intellect in selecting key characteristics. It is likely that many, if not most, of the definitions
reviewed in the next section were devised in this fashion.
Finally, definitions often grow convoluted because researchers do not clearly distinguish
among causes, characteristics, and consequences of the phenomenon being defined. Indeed, as
Stallings (2005) points out, definitions are not intended to be a collection of causal statements.
Quarantelli (2005a, p. 333) similarly argues that researchers must separate the conditions,
characteristics, and consequences of disasters when developing definitions. The definitions
presented below have been selected from the original works to emphasize where possible each
author’s statement of characteristics.
DEFINITIONS OF DISASTER
Although an effort was made to gather as many formal definitions of disaster as possible, no
claim can be made that those presented here exhaust the record. Those selected for inclusion do
seem to be among the most visible definitions presented over the decades. Since the mid-1990s,
when Quarantelli began assembling groups of disaster scholars to discuss definitions, the task
has been made easier by volumes he assembled (Quarantelli, 1998a; Perry & Quarantelli,
2005). In choosing definitions, there was a sense of need to accommodate interdisciplinary
study, but also to focus on the issue of disaster as a principally sociological construct.
Several classes of definitions are not included in this discussion. First among these are
mandated definitions that are generated as a matter of social or government policy. These are
usually used in making decisions about official disaster declarations or resource allocations
connected with mitigation, preparedness, response or recovery. The purposes for which such
definitions are devised are manifold, but not within a social scientific context. There are at
least two excellent discussions of mandated definitions in the recent literature (Britton, 2005;
Buckle, 2005). Similarly, hazards are not disasters and hazard-related definitions are included
What Is a Disaster?
5
only to the extent that they explicitly address the occasion of disaster. Also eliminated are
phenotypic definitions that focus on the surface features of an agent, such as natural versus
man-made.
The simple presentation of definitions of disaster also raises a challenge. Chronological
time, especially publication dates, is not a particularly effective ordering devise, as it implies
serial or sequential development. In practice, many people used a definition for years without
publishing it; some never wrote it down or published it only after using it implicitly for years.
Many researchers simply adopted another scholar’s definition, again explicitly or implicitly.
One remedy to this problem is to group definitions by era, with a simultaneous concern for
what might be called “paradigm” or “orientation.” Certainly the definitions proposed and the
studies conducted in the first decade of modern disaster research (1950s) influenced most of
the work that followed. But this approach must be tempered to acknowledge that definitional
foci have varied over the years. This condition sometimes places the same scholar in different
categories at different times. The imperfect solution adopted here is to examine three focal
areas: the classic approach and its variants, the hazards-disaster tradition, and the explicitly
socially focused tradition. Like all stage or sequence models, these three “traditions” can be
seen to overlap in time and to a small extent in content. They are acknowledged to be analytic creations designed to facilitate discussion. There is no guarantee, however, that different
observers might not place specific definitions in different places, or for that matter, devise
more or fewer categories. Nonetheless, as artificial ordering devices are concerned, they are
practicable.
THE CLASSICAL PERIOD AND ITS EVOLUTIONS
The classical period may be seen as beginning the end of World War II and closing with the
publication of Fritz’s definition in 1961. The influence of the thinking and writing in this period
on definitions of disaster , of course, extends to the present day. Three important intellectual
and research activities operated early in this period. Studies were conducted of the impact of
bombing on European and Japanese cities. The studies from Europe (United States Strategic
Bombing Survey, 1947; Ikle, 1951) were systematic and included the reaction of the population
as well as the customary examinations of physical damage. In 1951 and 1952, the National
Opinion Research Center at the University of Chicago conducted a series of studies of eight
disasters (mostly airplane crashes, but also fires and an earthquake). Charles Fritz oversaw the
NORC studies and the field teams included E. L. Quarantelli. The third development was the
formation of the Disaster Research Group, in 1952, at the National Research Council under
the auspices of the National Academy of Sciences (NAS-NRC). This group was charged with
conducting a review of the state of disaster research and conducted what has become a classic
series of studies (Williams, 1954).
Many of these studies left the meaning of disaster implicit, but the definitions that did arise
mentioned an event as catalyst for what now would be described as a failure of the social system
to deliver reasonable conditions of life. At a minimum, the data from these studies collectively
formed the first systematic (as opposed to journalistic or historical) information about human
behavior in disasters. It is important to make two observations about this era. First, while
the definitions explicitly mentioned an agent as catalyst (hence the use of the term “event”),
most really dealt with the social disruption attendant to the cause rather than the cause or agent
itself. Fritz’s (1961b) research on the therapeutic community that arose following disasters is an
important example of this emphasis on the social. It is easy to criticize these definitions as event
6
Ronald W. Perry
centered if one has not actually read and appreciated the human and social variables that were
actually studied. Second, the seeds of emergent norm thinking were sown during this period.
This framework was ultimately developed by sociological social psychologists (particularly
students of Mead’s symbolic interactionism), and influenced students of collective behavior
(particularly those interested in crowd behavior) and disaster researchers. It was manifest on the
definitional side among disaster researchers in the vision that social interactions were supported
by norms that might be rendered ineffective by disasters, thereby requiring different norms
until the environment began to stabilize again (Gillespie & Perry, 1974). Research following
this premise included Anderson’s (1969) study of change after the 1964 Alaska earthquake
and much later, Stallings’ (1998) presentation of “exceptions” and “exception routines” as a
perspective on disaster and the social order. Thus, although much of it was not published in
the open literature, this era saw a great deal of inductive research, some deductive research,
and much thinking that spawned attempts at theory development later. In effect, this period
generated the first real “database” for subsequent research and theorizing.
In this active research context, three formal definitions of disaster were published.
Anthony F. C. Wallace (1956a, p. 1), in a paper originally given as a committee report to
the National Academy of Sciences—National Research Council Disaster Research Group in
1954, characterized disasters broadly as situations that involve not just impact, but the threat
of “an interruption of normally effective procedures for reducing certain tensions, together
with a dramatic increase in tensions.” The social readjustment following these interruptions
was also cited as part of the definition of the disaster. This early definition is generic and reflects the general opinion of then contemporary disaster researchers that disasters were events
with negative social consequences. The use of the term “extreme situations” prefaced the later
concern that disasters may actually be a subcategory of a larger class of events. At about the
same time, Lewis M. Killian (1954, p. 67) proposed that disasters disrupt the social order,
producing physical destruction and death that becomes important because people must cope
by departing “from the pattern of norm expectations.” Killian retained the negative dimension
as a key feature of disasters as well as the importance of social consequences generated by
a need to change normative behaviors. Harry Estil Moore was associated with the Disaster
Research Group for some years, generating in the early 1960s what are now classic studies of
warning response behavior. As part of his studies of tornadoes in Texas, Moore (1958, p. 310)
also emphasized that a defining feature of disasters is that they make people adopt new behavior patterns; however, “the loss of life is an essential element.” These three definitions are
remarkably consistent with one another. Each characterizes disaster in terms of the impact on
social order, and each focuses on negative consequences. Emergent norm thinking is implicit
in all: the pattern of interrupted stability, followed by adaptation to the interruption, followed
by a resumption (though not necessarily unchanged) of behavior in a stable period. These
definitions also share a general or generic quality.
Fritz, working for the most part in the same tradition and on many of the same projects
as the first three authors, proposed a definition of disaster in 1961 (and reiterated it in 1968)
designed to capture the sociological notion of disaster. Fritz saw disaster as an event impacting
an entire society or some subdivision and including the notion of real impact with threat of
impact, but emphasized that “essential functions of the society [are] prevented” (1961a, p. 655).
This definition does not depart radically from the previous ones, but it attempts to be more
“precise” and detailed. It did specify disaster as an “event” that later critics would argue moved
the focus from strictly social and it also explicitly added “time and space” qualifications that
one might argue limited disasters to being rapid onset events, although that implication was
already implicit in the other definitions. There was also the rather strenuous requirement that
What Is a Disaster?
7
a “society or relatively self-sufficient subdivision” be affected. This is interesting because at
the time the definition was proposed (and since for that matter), little research was directed at
disasters affecting an entire society. For decades later, it appears that the liberal determination
of “relatively self-sufficient subdivision” allowed disaster researchers to embrace the definition
while studying communities and groups smaller than communities.
Fritz’s definition then was generated out of the intellectual context of the major disaster
research efforts of the 1950s and the social context of the cold war. The apparent societal
and governmental concern regarding a Soviet threat of an external attack came to be reflected
in the notion that disasters were both events and external to a focal society or social group.
In retrospect, one advantage of the definition was that it seemed to provide an umbrella for
much of the increasing number of studies done by a growing multidisciplinary body of disaster
researchers. After its publication, for decades many researchers simply adopted it verbatim or
pointed to it. Wettenhall’s (1975) studies of bush fire disasters; Perry, Lindell, and Greene’s
flood research (1981); and Perry’s study of a nuclear power plant accident (1985) are only
a few of many examples of those who adopted Fritz unchanged. Still, into the 21st century,
researchers pose definitions that embrace the basic tenets of Fritz work. Buckle (2005, p. 179),
speaking of a consensual definition of disasters, that one draws a sense of significant, irreversible loss and damage from disasters, requiring “the need of long term recovery.” Similarly,
Smith (2005, p. 301) proposed that disasters are events that produce death and damage and
cause “considerable social, political and economic disruptions.” Even Kroll-Smith and Gunter
(1998, pp. 161–163), who clearly consider their thinking not part of “classical disaster sociology,” describe incidents to be studied as disasters that largely meet the criteria in Fritz’s
definition; their argument is more about how and whom to study.
As recently as 2003, Henry Fischer, in accepting the Fritz definition, pointed out that
sociologists really study social change under disaster conditions (2003, p. 95). Like Fischer,
researchers began to accommodate slight variance from the original definition in what they
were studying by adding modifiers to the definition. Thus, over time, small changes began
to creep into the Fritz definition, introduced by researchers who largely embraced what they
believed was Fritz’s meaning. Four examples show this trend lasting well into the 1980s.
Gideon Sjoberg (1962, p. 357) characterized disaster as a “severe, relatively sudden, and
frequently unexpected disruption” of a social system resulting from some precipitating event
that is not subject to societal control. Thus Sjoberg introduces the notions that the precipitating
event is sudden onset, external to the system and not subject to control. On the surface, this
approach appears to tie disasters to the state of technology that might define control, but
as Mileti (1999) indicated much later, humans can exert control in some cases by simply
changing their settlement patterns. In the same year, Cisin and Clark (1962, p. 30) appeared
to drop some of Fritz’s qualifiers by saying a disaster is any event that “seriously disrupts
normal activities.” In elaboration, these authors added the explicit qualifier that the disaster
may result from a threat that does not materialize as well as from an actual impact. This adds a
new dimension to potential disaster studies (threats of destruction or disruption), while at the
same time introducing some latitude in the stringent target of disasters set by Fritz by noting
the disruption can be of “normal activities” and not specifying the social system.
Barry Turner (1978, p. 83) re-created part of the Fritz definition in defining disaster, but
emphasized the notion that there must be a collapse of social structural arrangements that
were previously “culturally accepted as adequate.” Turner’s definition was given in the context
of a book on disasters with origins in human forces (“man-made”) and adds the notion that
disasters take place when precautions that are culturally based fail to allow continuation of
“normal” behavior patterns. Drabek (1986, p. 7) adopted Fritz’s words verbatim but prefaced
8
Ronald W. Perry
the definition with the provision that “disasters are accidental or uncontrollable events, actual
or threatened.” The provision includes the notion of threat as a disaster, as well as clarifies the
possibility of accidental origin (as Turner also probably sought to do). All these definitions
represent not so much “drift” from Fritz’s conception, but as adding qualifications to the
definition that ultimately formalized an expansion of the phenomena that could be studied as
disasters.
If one tries to trace the definition proposed by Fritz into contemporary disaster research,
it appears that the present evolutions share in common a focus on the social order that Fritz
and the researchers before him saw as a key defining feature. While the authors cited below
may or may not see themselves as operating in a “classical disaster” context, their definitions
do reflect a concern, sometimes implied, with the key defining features mentioned by Fritz. A
feature that distinguishes each from Fritz, however, is an explicit emphasis on disasters and
social process or change and the notion that disasters may be a category of some larger class
of events. Perhaps Gary Kreps (1998, p. 34) remains closest to Fritz when he defines disasters
as “non-routine events” that create social disruption and physical damage. In elaborating his
definition, he focuses on four key defining properties—forewarning, magnitude of impact,
scope of impact, and duration of impact. This definition and elaboration constructs disasters
as a category of events within combinations of the defining properties; it reflects the work of
Barton (1963, 1969) and prefaces Kreps’ later work on disaster taxonomy (1989).
Robert Stallings created a picture of disasters that firmly placed them within a context
of classical social theory, while at the same time emphasizing the notions of disruption and
change. Stallings (1998, p. 136) examines routines, exceptions, and exception routines: the
social order is seen as routinization and “Disasters are fundamentally disruptions of routines.”
Stallings also acknowledges in his formulation that disasters are only one kind of occasion that
interrupts routines in social life. Later, Stallings (2005, p. 263) defined disaster as “a social
situation” precipitated by nonroutine destruction by forces of nature. Stallings was writing in
the context of natural disasters and undoubtedly did not intend to limit disasters to agents of
the natural environment. Stallings’ work is important both for its extension of Fritz’s definition
and because he firmly places disaster in the social order. In this latter vision, his efforts toward
definition are also compatible with the definitions in the subsequent discussion of disasters
as largely social phenomena. Boris Porfiriev (1998b, p. 1) also sees disaster as an event that
destabilizes the social system, indicated by a failure of normal functioning that requires an
intervention to reinstate stability. Again, one sees an emphasis on disaster as transition or
change that involves vulnerability and requires different patterns of social intercourse.
The spirit embodied in Fritz’s definition is certainly reflected in these definitions; there
is an element in each that appears to retain the “event” perspective. As an observer, I also feel
that they are substantively different in that they explicitly (in the definition or in each author’s
elaborations) emphasize process, adaptation, and change. These notions were more implicit
in the approach taken by Fritz. Each of these authors seems to not just recognize, but also to
emphasize, a cycle of stability–disruption–adjustment that characterizes disasters.
THE HAZARDS-DISASTER TRADITION
Another tradition of viewing and defining disasters grew out of the hazards perspective common
in the literature of geographers and other geophysical scientists. As Quarantelli (1998b) has
pointed out, a hazards perspective focuses on the hazard—earthquakes, tornadoes, floods, and
so forth—and understanding it. Although there may be a concern with social and other issues,
What Is a Disaster?
9
the real emphasis is on the processes associated with the target agent. One could probably find
earlier statements, but the classic elaborations of the hazard approach are usually taken to be
those of Burton and Kates (1964) and Burton, Kates, and White (1978). Within this context
generally, a disaster is viewed as an extreme event that arises when a hazard agent intersects
with a social system (“the human use system”). Technically, then, disasters are events that
take place as part of normal environmental processes; they are not the principal focus of
study. It is often pointed out that a hazards approach is a legitimate focus of study but that it
is different from the classic sociological approach to disasters. From a hazards perspective,
Quarantelli (2005, p. 342) argues that hazard cycles and agents are the focus, making disasters
epiphenomena. Indeed he points out that some phenomena studied legitimately as disasters have
no identifiable originating agent (Quarantelli, 2005a, p. 347), such as famines and computer
system failures. This is not to say that researchers operating within a hazards framework have
not generated valuable findings relative to human behavior in disasters and it does create
an opportunity to examine definitional strategies different from the classic disaster research
approach.
John Oliver (1980, p. 3) defined disaster as a part of the environmental process that is
of greater than expected frequency and magnitude and causes major “human hardship with
significant damage.” The classical era is clearly reflected in this definition, but the critical
hazards issue of a cyclic environmental process is also present. Susman, Okeefe, and Wisner
(1983a, p. 264) are closer to the traditional geographers’ view when they define disaster as
“the interface between an extreme physical event and a vulnerable human population.” Hewitt
(1998, p. 77) elaborates a view of disaster as events in which “physical agents define the
problem.” In 1983 he argued that disasters may be seen as unexpected and unprecedented
impacts that “derive from natural processes of events” (Hewitt, 1983a, p. 10). Each of these
definitions highlights the traditional concern of hazards researchers with the cycle of hazard
agents in their vision of disasters.
Recently, hazards researchers studying disasters have moved slightly from what might be
considered an “agent centered” approach to a greater focus on vulnerability. David Alexander
(1993, p. 4) pointed out that natural disasters can be thought of as quick-onset events with
significant impacts on the “natural environment upon the socio-economic system.” In later
writing, he elaborated this by saying that disasters are not defined by fixed events “but by social
constructs and these are liable to change” (Alexander, 2005, p. 29).” The concern expressed by
Alexander is that disasters are not just the events but also the social consequences (which are
ever changing) of the event. Dennis Mileti (1999, p. 3) also emphasizes that disasters flow from
overlaps of the physical, built, and social environments, but that they are “social in nature.”
Mileti emphasizes that humans can be seen as creating disasters through their encroachment on
the physical environment. Although he still places the origins of disasters in a hazard context,
Mileti is explicit about the social emphasis when studying the events. Finally, and most firmly
in a vulnerability context, Susan Cutter (2005b, p. 39) argued that the issue is not disasters as
events but instead human “vulnerability (and resiliency) to environmental threats and extreme
events.”
Each of these definitions retains the hazard origins of disasters, but also moves to examine them in social terms, particularly of vulnerability and resilience. As Quarantelli (2005a,
p. 345) indicates, this emphasis reinforces the traditional notion that in defining and studying
disasters, one should look first at social systems, since they (not the agent) are the real source
of vulnerability. To the extent that the researchers in a hazard-disaster tradition are moving
in this direction, they are converging with sociological researchers to place people and social
relationships at the core of disaster study.
10
Ronald W. Perry
DISASTERS AS A SOCIAL PHENOMENON
Finally, although still in the tradition of the original disaster studies, there has been a group who
explicitly focused on social phenomena as the defining feature of disasters, within the context of
social change. At least some of these researchers were active during the classical era and nearly
all would place their intellectual roots in that time period. Indeed, these definitions are similar
to those of Kreps (1998), Stallings (1998), and Porfiriev (1998b) whom I have placed as later
evolutions of the classic disaster era. However, these conceptions of disasters are distinct in
their emphasis on social phenomena, their attention to vulnerability as socially constructed, and
the idea of social change, all to the near exclusion of physical agents. In the latter feature, these
definitions depart from both the classical era derivations and the hazard-disaster perspective.
Allen Barton (1963, 1969, p. 38) saw disasters as one collective stress situation arising
when members “of a social system fail to receive expected conditions of life from the system.” Barton then moved to a classification scheme that created a matrix of four dimensions
(scope of impact, speed of onset, duration of impact, and social preparedness) and proceeded
to characterize events in the cells in social and interpersonal terms. In 1989, Barton reminded
colleagues that the bulk of disaster studies focus upon “events at a community level caused
by physical agent” (1989, p. 348). He wanted to emphasize that, because we are sociologists,
there was a need to define our subject matter more firmly in the realm of the social. Subsequently Barton (2005) revised the dimensions of his typology to address the cross-classified
dimensions of scope (national, regional, segmental, local) by concentration in time (sudden
gradual chronic). Again, his discussion of what belonged in the cells of his matrix described
the social dimensions of events, not the events themselves (which he cited only as examples).
Some would say Barton evaded defining disasters except to call them a category of collective
stress situations and then to describe many different classes of this category. Another way of
reading the work is to see Barton’s classes (the matrix of scope by time) as many specific types
of disaster—without specifying a label for each—created in social terms.
E. L. Quarantelli’s career spans the classical era through the present. He was involved in
the early research efforts, conducted much research himself, co-founded (with Russell Dynes)
the Disaster Research Center in 1963, and trained generations of disaster researchers. By 1966
he had begun, like Barton, to publish typologies for disaster research (Quarantelli, 1966). He
(Quarantelli, 2000, p. 682) identifies disasters in terms of a variety of defining features. They:
(1) are sudden-onset occasions, (2) seriously disrupt the routines of collective units, (3) cause
the adoption of unplanned courses of action to adjust to the disruption, (4) have unexpected life
histories designated in social space and time, and (5) pose danger to valued social objects. He
subsequently emphasized that disasters represent vulnerability, reflecting “weaknesses in social
structures or social systems” (Quarantelli, 2005a, p. 345). In this characterization, Quarantelli
emphasizes neither an event nor a physical place or time as relevant to disasters. Instead, the
entire conception is social: vulnerability is socially constructed by relationships in the social
system and disasters are based in the notion of social changes. The definition may have roots
in the classical era research, but clearly departs in significant ways.
While Quarantelli has used this definition for decades—as well as advocated it as a
model—one can also trace a convergence to his point of view in the literature. One early
example is Kai Erikson’s (1976, p. 254) view of disasters as sudden causes of harm to the
physical and social environment but with an emphasis that they “are socially defined as having
reached one or more acute stages.” While an agent is implied here, the focus is on a social
definition and vulnerability that might be modified (through social change). Lars Clausen
(1992, p. 182) emphasized the latter, arguing that disasters flow from normal social change even
What Is a Disaster?
11
though their consequences are negative and their frequency rare. The reference to normality
underscores the point that vulnerability lies within the social structure itself and is a regular part
of human intercourse. Similarly, Gilbert (1998, p. 13) argues that “disasters are not a function
of agents, but are social in origin;” like Mileti, he sees disaster as stemming from humaninduced vulnerability. Parenthetically, David Alexander and Susan Cutter, who work in the
hazards tradition, emphasize social vulnerability and change when they address disasters.
Russell Dynes (1998, p. 13) also fits within this tradition, defining disaster as occasions
when norms fail, causing a community to engage in extraordinary efforts “to protect and
benefit some social resource.” Even closer to Quarantelli’s view is Rosenthal’s (1998, p. 226)
discussion of disaster as a socially defined occasion, related to social change that is “recognized
across social time as a radical change” in the normative environment. The reference to social
time particularly sets this definition apart from most others. It is also interesting that Rosenthal
and Quarantelli have each stressed the need to develop an overarching category that contains
disasters (the beginnings of typology) and that each chose the term crisis for that category.
There is also an approach to defining disasters socially that arose among researchers and
others interested in cross-national or cross-cultural aspects of the phenomenon. For example,
Bates and Peacock (1993, p. 13) characterize disasters as a social event arising out of “a process
that involves a socio-cultural system’s failure” to protect its population from external or internal
vulnerability. The event notion has crept into the definition, but for these authors, disasters are
social phenomena that have roots in the social structure itself. On the other hand, Jigyasu
(2005) bases disasters in the social, but his almost metaphysical view of the phenomenon
is not clearly tied to the social structure, being rather an intellectual state. Conversely, for
Horlick-Jones (1995, p. 311), “disasters are disruptions in cultural expectations” that result in
the perception that institutions cannot keep hazards in check. He points out too that disruptions
stem from the ways in which society deals with vulnerability. Dombrowsky (1998, 2005)
follows this approach by relating disasters to knowledge. His view is that disaster is the collapse
of cultural protections—captured in habits, folkways, laws, or policies—that either deflect or
fail to deflect the threatening forces to which societies are exposed. For Dombrowsky, the
disaster is social; it is engendered in social structure and can be attacked only via that route.
Finally, the anthropologist Anthony Oliver-Smith (1998, p. 186) sees disaster as an event
that combines destructive agents with a vulnerable population disrupting “social needs for
physical survival, social order and meaning.” While Oliver-Smith includes the words event
and process, the definition is social, placing the disruption and the vulnerability each within
the social structure. Interestingly, Oliver-Smith and Bates and Peacock have studied disasters
in developing countries and their definitions and research link disaster (and development) to
social change. Finally, Arjen Boin (2005, p. 159) believes that disasters flow from the normal
functioning of social systems that take place when the “life sustaining functions of the system
break down.” Boin, like Barton, Quarantelli, Kreps, and Stallings, argues that disasters are a
subclass of a larger class; Barton called the larger type collective stress situations, while Boin,
like Quarantelli and Rosenthal, uses the label crisis. For Boin, disasters are rooted in social
structure and changes that cause disruption.
In closing, although interdisciplinary in their training and international in origin, these authors share a conception of disasters that places the phenomenon firmly in social relations. The
disaster is characterized as a social disruption that originates in the social structure and might
be remedied through social structural manipulations. Further, social structure can be seen as social change “analytically frozen at one point in time” (Quarantelli, 2005a, p. 340). This means
that each of the preceding definitions hinges upon social change. It is the combination of these
two features that distinguishes these definitions from others more rooted in the classical era.
12
Ronald W. Perry
CONSENSUS ON DEFINITIONS
In almost every definition cited in the foregoing section, the author or authors included an
elaboration to explain intent and often demonstrated causes and consequences of disasters.
These elaborations contain important messages, but space limitations prohibited their presentation here. The following comments on themes often rely on these elaborations as well as my
interpretation (right or wrong). In the end, the references are there to be checked by skeptics.
More than three dozen definitions of disaster have been presented in this chapter. It would
be unrealistic to expect to find homogeneity among them. But clearly there are similarities
and overlap; it can certainly be argued that the three artificially constructed “families” of
definitions show considerable similarity within groups. And one would not expect to find
common definition in any professional grouping, perhaps especially among social scientists.
It is possible, however, to assess levels of general consensus across the definitions as a means
of inferring agreement about what disaster researchers see as their field of study.
In this regard, the degree of consensus depends both upon the observer and on the level
of specificity demanded to define consensus. Quarantelli (2005a, p. 338) summarizes his
assessment by observing that “it would be difficult to deny that there is a substantial lack of
consensus” about the meaning of the term disaster. I agree that comparing the detail of each
definition (except when multiple authors adopt verbatim the definition of another author) yields
an environment of significant differences. Similarly, there are differences in social scientific
orientation as well; compare the positivist approaches of Stallings, Kreps, and Dombrowsky
with the more interpretive approach of Kroll-Smith and Gunter, versus the almost mystical–
phenomenological approach of Jigyasu. At the same time, the task becomes more manageable
if the goal is to identify common themes in the definitions.
In discussing what he calls the current paradigm of disaster research, Quarantelli (2005a,
p. 339) points out that it is rooted in two fundamental ideas. First, disasters are inherently
social phenomena. It is not the hurricane wind or storm surge that makes the disaster; these
are the source of damage. The disaster is the impact on individual coping patterns and the
inputs and outputs of social systems. Second, the disaster is rooted in the social structure and
reflects the processes of social change. It is from these features of the social system that we find
vulnerability to the particular source. In effect, this vision of the field is reflected in the majority
of the definitions reviewed here. Looking for themes is a fruitful way of capturing concepts
of the field that might be obscured in the specific language and detail of a comparatively
short definition. Of course there is the risk of misinterpretation when making inferences about
themes, but social science is filled with such risk and at some point it is more irresponsible
to say nothing than to risk being wrong. Kaplan (1964) warned about reconstructed logic (the
scientist’s “cleaned up” reconstruction of what they do) versus logic in use (what an observer
would see a scientist do). Definitions can be seen as a form of reconstructed logic, and by
identifying themes one is at least attempting to capture the logic in use.
I view a theme as arbitrarily specified as a common opinion by many (not even most) of
the authors of the definitions reviewed. Studying disasters means you look for what? There is
wide agreement (outside the classic hazard perspective) that disasters are social, that they are
understood in human interaction. The researchers captured here under the rubric of “disaster
as a social phenomenon” thereby often use the word occasion rather than event when speaking
of disaster. There is also wide agreement that in disaster one finds disruption of the social.
Some definitions and elaborations mention the source of the disruption as an event or force,
but almost all agree on the social fact of disruption and that people’s lives are being disrupted.
Many agree that disasters stem not from the agent that causes the disruption, but from the
What Is a Disaster?
13
social structure of norms and values, hence the protections. Vulnerability, a part of many of
the definitions, is to be found in social structure and disruption is the outcome of vulnerability.
There is some consensus, by inference, that the magnitude of a disaster should be measured
not in lives or property lost, but by the extent of the failure of the normative or cultural system.
Another fairly common theme is the issue of resilience. Some definitions in the classical
tradition mark the end of one phase of disasters as the point at which normative stability is
restored, while others call this restoration the implementation of emergency measures (norms)
or exception routines. The link to emergent norm thinking is unmistakable. Typically, those
who emphasize vulnerability include the notion of resilience in some form. Finally, although
some authors speak of disasters as social problems, there is a general consensus that disasters
are best understood in a context of social change. Carr seems to have originated this thinking
in 1932 and it is present in much of the work of the classical era as well as being a staple of
those who define disasters as exclusively a social phenomenon. Among the latter, and in some
of the recent hazard-disaster definitions, as well as a handful of the definitions that evolved
from the classical period, there is an emphasis on defining disasters in social time and space
rather than physical time and space. As yet, more disaster researchers ignore social time and
space than understand it or incorporate it into their research.
While all these common themes fit well within Quarantelli’s exposition of a current
disaster paradigm, much disagreement about disasters as an area of study remains. Some of
the disagreement about disasters rests in issues that are not exclusively definitional. That is,
there is disagreement about how disasters should be studied, how the definitions proposed
by different groups (citizens, policy formulators, etc.) should be treated, the nature of social
science, even whether disaster research is subsumed by social science, as well as disciplinary
differences such as the hazards-disaster distinction. A few differences are based on definition
and relate to such issues as the extent to which disasters originate outside a social system, the
degree to which social change is emphasized, the centrality of the role of an agent, and how
disaster consequences are to be conceived.
Some of the disagreement about disasters seems to stem from what are really taxonomic
issues or at least from the typologies or classifications that are produced by taxonomic thinking
(Perry, 1989). These are essentially disagreements about what kinds of characteristics should
be included in the definition of a disaster and are expressed in different ways. Many of the
scholars who authored the definitions have noted that disasters seem to be part of a “larger
class of events.” Indeed many who have proposed definitions from across the three perspectives
included with their definitions a set of dimensions—such as social preparedness, speed of onset,
scope and duration of impact—to create categories of disasters. Others have talked about how
one can distinguish disasters from events that “look like” disasters. Quarantelli (2005a, p. 333)
distinguishes disasters, catastrophes, and crises. Boin, Stallings, and Rosenthal have likewise
separated disasters and crises (although using different referents for the latter term). Stallings
(1991) and Quarantelli (2005a, p. 336) have proposed that situations involving conflict belong
in a category different than “disaster.” Similarly, Quarantelli (2005a, p. 335), as well as the
authors of several of the definitions reviewed in this chapter, suggest the elimination of slow
developing and diffuse events from the category of “disaster.”
Quarantelli (1987a) makes a most convincing case for investing effort in taxonomy to
create meaningful classification systems. He points out that the many empirical studies of
“disasters” have begun to produce anomalous findings; using only one example, we know
that serious mental health consequences, rare in most studies based on floods, tornadoes,
hurricanes, and earthquakes, appear to be greater in cases associated with conflict situations
(see Perry & Mankin, 2004 for a discussion of terrorist attacks). One explanation for such
14
Ronald W. Perry
anomalies is classification error, comparing two things that are similar in phenotype, but are
different genotypes. Classification systems are a way of sorting occasions and findings to make
appropriate comparisons based on genotype. Quarantelli argues that disaster researchers need
a classification system based on general dimensions that not only distinguish among different
disaster agents, but also specify differences within one category of agent (1987a, p. 26). Drabek
(1986, p. 6) stressed nearly 20 years ago that taxonomy is the “most pressing issue confronting
the field at this time.”
Certainly many disaster researchers have felt this need. Proposals for dimensions for
classification schemes have historically accompanied efforts at definition since the earliest
days. Two comprehensive typologies have been devised. Barton (1963, 1969, 2005) created
a host of categories in a typology of collective stress situations, and Kreps (1989) devised an
intricate system by looking at domains, tasks, resources, and activities (DTRA). There have
consequently been many varied attempts to start disaster research down the taxonomic path,
but with mediocre success. For the most part, those who conduct and interpret disaster research
have neglected existing typological systems and rarely have chosen to qualify their findings in
terms of the dimensions that are common in the literature: speed of onset, scope, and duration
of impact and the like. The confusion and apparent anomalies that derive from this practice are
likely to continue until researchers begin to document such qualifications or to operate within
some typology. The explosion of disaster studies described at the beginning of this chapter
will only exacerbate the problems.
The real challenge and danger is for the growth of disaster research as a field of study. As
Hank Fischer indicated, most disaster research is not about the meaning of disaster. Descriptive
studies can be (and long have been) generated with little attention to issues of theory or
paradigm for that matter. However, as Drabek (1989) has warned, the creation of models and
the production of viable explanations, predictions, and efforts at control move well beyond the
descriptive task. To assemble a meaningful body of knowledge about disasters (Perry, 2005),
it is absolutely critical that disaster researchers pursue Quarantelli’s admonition not just to
develop greater consensus regarding the meaning of the term, but also begin scrupulous use of
typologies. Failing that, the field will continue to amass a disconnected collection of descriptive
research that cannot be linked via existing conceptual tools.
AN AGENDA FOR FUTURE RESEARCH
The critical issues raised in this chapter do not focus on further research. The call here is
for thinking, not for more doing. Research can and will continue, but sociologists must renew and revitalize their focus on conceptual matters. As indicated in the previous section,
there are already hundreds of studies of individual, organizational, and institutional behavior
in the literature that describe, and in a few cases attempt to explain, actions during times of
“disaster.” The problem is that a variety of views co-exist—some differing significantly—
of the defining features of disaster. As Quarantelli and others have argued, the research
record has accumulated under these varying definitions of disaster and has begun to produce apparently conflicting findings, when there may or may not be real differences. The reports of “looting behavior” in New Orleans following Hurricane Katrina (Quarantelli, 2005b)
again underscore that more attention needs to be paid to the social context of the behavior (e.g., conflict or abandonment) than just to its characterization as taking place following a “disaster.” Indeed, the problem of “what is a disaster” will never be solved by more
fieldwork.
What Is a Disaster?
15
The real work to be done with respect to definitions of disaster has to do first with
conceptualization; one needs to decide what disaster means. More specifically, each researcher
needs to decide. This is not an empirical task. One must decide on fundamentals such as whether
disasters are social phenomena or are the events with which they are often associated or even
some natural or technological process. The stream of definitions recounted in this chapter seeks
a starting point by identifying areas of consensus in what might be seen as a sea of differences.
The differences count too. A significant point is that this practice of making explicit our
definition of disaster has begun to take hold, although there is much more to be done.
The second part of the work to be done focuses on dialog among sociologists and disaster
researchers. It is no longer appropriate to expect that a researcher can continue to do studies
without specifying what constitutes a disaster. Further, the task of defining disasters should
no longer be treated as an unnecessary abstraction that occupies the minds of a few senior
(old) disaster researchers. There needs to be a serious engagement on the definition issue. A
concern with taxonomy—the reasoning that underlies typologies or classification systems—
should logically evolve out of this dialogue or engagement. Clearly, as other disciplines (such
as botany and zoology) have found, we need to further specify our subject matter. This work
too has seen a modest start. Taxonomic thinking cannot be characterized as the “easy work.” It
demands that one carefully understand the growing field of findings, appreciate the meaning
of disaster in conceptual terms, and engage in both inductive and some deductive reasoning
to support the creation of classification systems. The plural use of systems is an operative and
indicative term here. There need not be a single typology; many can coexist. But there must be
one or some typologies and they must be widely scrutinized by the disaster community. The
more scrutiny, the more likely and quickly consensus at some level will begin to emerge.
Perhaps most critical, researchers will need to characterize their ongoing research in terms
of one or more typologies. “Where does this study fit in the disaster cornucopia?” should be
asked with each piece of research. At the same time, a need arises to consider the disaster
findings of days (decades) past. When we cite those findings we must begin to group them into
one or more of today’s typologies. In this way, it would be routine to separate findings about
looting behavior in situations that do and do not involve conflict. One practical outcome of the
use of typologies is that we can reduce the potential ambiguity associated with interpreting
our findings across events and at the same time present a clearer and more precise picture to
those who may be using our findings to devise social policy. We will never “research” our way
out of this problem. Such a tactic will only bury the field further in a kind of intellectual and
conceptual muddle that will produce obfuscation and confusion.